# cbloom rants

## 1/15/2006

### 1-15-06

Here's a proposal for a new way to do betting odds. In simple win/loss bets, you can either bet on side A or side B. You can put whatever amount of money you want on each side. By putting money on a side, you are buying shares in that side equal to the % of the total amount of money put on that side. When the outcome is determined, the winning side gets all the money, and it is split based on the % of money put up.

For example, pretend this was used on the Colts-Steelers game. Lots of people bet, and when you go to the action there's \$1000 on the Colts and \$500 on the Steelers. You bet \$5 on the Steelers. Now the Steelers win, so they get the full \$1505 pot. You get 5/505 of that = \$14.9. Basically the odds are set by the amount bet on each side. This is an auto-equalizing system where if there is any imbalance, the bettors will fix it by betting on the more favorable option.

This is particularly nice in multiway action, for example betting on the team to win the superbowl. Rather than giving odds for each team, the bettor can simply put their money on the team they like to win. The odds are set automatically based on where the money goes.

One problem with this system is that it's susceptible to manipulation. Various things could be done to prevent that. The simplest would be to make the betting blind. You simply get to pick one side and don't get to see how much money is placed on each side.

Clearly betting N on one side should be no different than betting N/2 twice. One way to do this is for the odds of each bet to be set based on the final amount in each pot. If you put in a big bet at the last minute, it hurts your odds, so that's no exploit. Another interesting way is for bets to be added one dollar at a time, with each dollar going it at the odds for that dollar. So, when I bet \$5 on the Steelers in the example above, really I'd get \$1 at 1000:501 and \$1 at 1000:502 and \$1 at 1000:503 etc. This is actually more favorable for the bettor than getting just the final odds.